The CDK is a Java library for cheminformatics and is employed in a variety of applications. Given that many problems in cheminformatics involve analysis of collections of molecules, accessing cheminformatics in a statistical modelling environment enables flexible modelling and analysis of chemical information. R is an open source statistical modelling and programming environment and provides extensive support for data munging and machine learning, but does not support cheminformatics natively. As a result, a number of packages have been developed to provide such capabilities within the R environment. In this talk we discuss the rcdk, an R package that makes the CDK accessible from within R, After a brief discussion of the design of the rcdk [1], we briefly survey the ecosystem of packages that has grown up around rcdk to enable chemical data science within R. We then describe how these efforts have fed into larger, originally third-party R-based projects that require cheminformatics functionality, using the example of RMassBank [2]. Building on the rcdk functionality, the automated cheminformatic and spectral curation workflow has enabled the release of over 16,000 open mass spectra to the open spectral library MassBank [3] - which itself surfaces the structural information to the public using CDK libraries. The seamless integration between rcdk and the CDK enables consistent cheminformatics functionality for end users and developers alike. These once independent projects have now grown into a (spontaneous) geographicaly distributed, collaborative development between the CDK, rcdk, RMassBank and MassBank teams, driven mainly by communications via Github, showing that the ability to combine open source projects leads to products that are more than the sum of their parts.

How the RCDK enables open source cheminformatics in R: From fingerprints to mass spectra

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.